Let’s say you’ve got a nice, digitized version of some scientific data, and you’ve already made reasonable choices about how close to the raw data you want to get in what you preserve. Better yet, you ...